953版本节点替换

查看集群状态

[gbase@gnode1 gcinstall]$ gcadmin

[gbase@gnode1 gcinstall]$ gcadmin showdistribution

模拟集群异常状态

[gbase@gnode1 gcinstall]$ gcadmin

[gbase@gnode1 gcinstall]$ gcadmin showdistribution

将节点变为不可用

[gbase@gnode1 gcinstall]$ gcadmin setnodestate 192.168.138.175 unavailable

删除被替换节点的feventlog

[gbase@gnode1 gcinstall]$ gcadmin rmfeventlog 192.168.138.175

查看节点状态

[gbase@gnode1 gcinstall]$ gcadmin

管理服务替换

[gbase@gnode1 gcware_server]$ cd /data/192.168.138.173/gcware/gcware_server/

[gbase@gnode1 gcware_server]$ ./gcserver.py --prefix=/data --host=192.168.138.175 --dbaUser=gbase --dbaPwd=gbase –overwrite

查看节点状态

[gbase@gnode1 gcware_server]$ gcadmin

注：175：gcware转换为open状态

调度服务替换

[gbase@gnode1 gcware_server]$ cd /home/gbase/gcinstall/

[gbase@gnode1 gcinstall]$ ./replace.py --host=192.168.138.175 --type=coor --dbaUser=gbase --dbaUserPwd=gbase --generalDBUse=gbase --generalDBPwd=gbase --overwrite --sync_coordi_metadata_timeout=3000

查看节点状态

[gbase@gnode1 gcinstall]$ gcadmin

注：coordinator管理服务状态由UNAVAILABLE转化为open

计算服务替换

拿到旧分布策略

[gbase@gnode1 gcinstall]$ gcadmin getdistribution 1  distribution_info_1.xml

注：1为distribution ID，使用gcadmin showdistribution命令查看distribution ID，当前为1

创建中间分布策略文件

[gbase@gnode1 gcinstall]$ cat distribution_info_1.xml

[gbase@gnode1 gcinstall]$ cp distribution_info_1.xml distribution_info_2.xml

[gbase@gnode1 gcinstall]$ vi distribution_info_2.xml

注：将故障ip：192.168.138.175从配置中删除

生成中间分布策略

[gbase@gnode1 gcinstall]$ gcadmin distribution gcChangeInfo_2.xml

查看新的分布策略

初始化并重分布中间策略

[gbase@gnode1 ~]$ gccli -ugbase -pgbase

GBase client 9.5.3.27.140fa78d1. Copyright (c) 2004-2024, GBase.  All Rights Reserved.

gbase> initnodedatamap;

Query OK, 0 rows affected, 5 warnings (Elapsed: 00:00:00.32)

gbase> rebalance instance;

Query OK, 1 row affected (Elapsed: 00:00:00.15)

gbase> select status,count(*) from gclusterdb.rebalancing_status group by status;

+-----------+----------+

| status    | count(*) |

+-----------+----------+

| COMPLETED |        1 |

+-----------+----------+

1 row in set (Elapsed: 00:00:00.05)

gbase>

注：检查进度，指导所有的都是completed状态

gbase> set global gcluster_rebalancing_concurrent_count=0;

Query OK, 0 rows affected (Elapsed: 00:00:00.01)

gbase>        

注：为防止并发过大，可以先关闭，并发值可0-15值，也可调整rebalance任务优先级，在跑批开始前，如任务还未完成，需设置并发为0，优先业务跑批；

计算节点替换

[gbase@gnode1 gcinstall]$ ./replace.py --host=192.168.138.175 --type=data --dbaUser=gbase --dbaUserPwd=gbase --generalDBUse=gbase --generalDBPwd=gbase --overwrite --sync_coordi_metadata_timeout=3000

查看节点状态

[gbase@gnode1 gcinstall]$ gcadmin

注：故障数据节点状态由UNAVAILABLE转换为open

查看分布策略

[gbase@gnode1 gcinstall]$ gcadmin showdistribution

注：如上replace操作，会自动新建一个最终策略3，其中已经将故障恢复的节点添加进去，冰洁最老的策略1已自动删除

初始化最终策略

开始重分布并监控进度

[gbase@gnode1 gcinstall]$ gccli -ugbase -pgbase

GBase client 9.5.3.27.140fa78d1. Copyright (c) 2004-2024, GBase.  All Rights Reserved.

gbase> initnodedatamap;

ERROR 1707 (HY000): gcluster command error: (GBA-02CO-0004) nodedatamap is already initialized.

gbase> rebalance instance;

Query OK, 1 row affected (Elapsed: 00:00:00.18)

gbase> select status,count(*) from gclusterdb.rebalancing_status group by status;

+----------+----------+

| status   | count(*) |

+----------+----------+

| STARTING |        1 |

+----------+----------+

1 row in set (Elapsed: 00:00:00.06)

gbase>

注：初始化报错可以忽略；为防止并发过大，可以先关闭，并发值可0-15值，也可调整rebalance任务优先级，在跑批开始前，如任务还未完成，需设置并发为0，优先业务跑批；

清理环境

[gbase@gnode1 gcinstall]$ gcadmin showdistribution

清理中间分布策略，可以看到本次中间分布策略ip为2

[gbase@gnode1 gcinstall]$ gccli -ugbase -pgbase

GBase client 9.5.3.27.140fa78d1. Copyright (c) 2004-2024, GBase.  All Rights Reserved.

gbase> refreshnodedatamap drop 2;

Query OK, 0 rows affected, 6 warnings (Elapsed: 00:00:00.61)

gbase>

[gbase@gnode1 gcinstall]$ gcadmin rmdistribution 2

查看状态

[gbase@gnode1 gcinstall]$ gcadmin

[gbase@gnode1 gcinstall]$ gcadmin showdistribution